3 require_once('Tokenizer.php');
5 if( $GLOBALS['wgUseWikiHiero'] ){
6 require_once('extensions/wikihiero/wikihiero.php');
8 if( $GLOBALS['wgUseTimeline'] ){
9 require_once('extensions/timeline/Timeline.php');
14 # Processes wiki markup
16 # There are two main entry points into the Parser class: parse() and preSaveTransform().
17 # The parse() function produces HTML output, preSaveTransform() produces altered wiki markup.
20 # objects: $wgLang, $wgDateFormatter, $wgLinkCache, $wgCurParser
22 # NOT $wgArticle, $wgUser or $wgTitle. Keep them away!
24 # settings: $wgUseTex*, $wgUseCategoryMagic*, $wgUseDynamicDates*, $wgInterwikiMagic*,
25 # $wgNamespacesWithSubpages, $wgLanguageCode, $wgAllowExternalImages*,
28 # * only within ParserOptions
31 #----------------------------------------
32 # Variable substitution O(N^2) attack
33 #-----------------------------------------
34 # Without countermeasures, it would be possible to attack the parser by saving a page
35 # filled with a large number of inclusions of large pages. The size of the generated
36 # page would be proportional to the square of the input size. Hence, we limit the number
37 # of inclusions of any given page, thus bringing any attack back to O(N).
40 define( "MAX_INCLUDE_REPEAT", 5 );
42 # Allowed values for $mOutputType
43 define( "OT_HTML", 1 );
44 define( "OT_WIKI", 2 );
45 define( "OT_MSG", 3 );
47 # string parameter for extractTags which will cause it
48 # to strip HTML comments in addition to regular
49 # <XML>-style tags. This should not be anything we
50 # may want to use in wikisyntax
51 define( "STRIP_COMMENTS", "HTMLCommentStrip" );
53 # prefix for escaping, used in two functions at least
54 define( "UNIQ_PREFIX", "NaodW29");
58 # Cleared with clearState():
59 var $mOutput, $mAutonumber, $mDTopen, $mStripState = array();
60 var $mVariables, $mIncludeCount, $mArgStack, $mLastSection, $mInPre;
63 var $mOptions, $mTitle, $mOutputType;
72 $this->mOutput
= new ParserOutput
;
73 $this->mAutonumber
= 0;
74 $this->mLastSection
= "";
75 $this->mDTopen
= false;
76 $this->mVariables
= false;
77 $this->mIncludeCount
= array();
78 $this->mStripState
= array();
79 $this->mArgStack
= array();
82 # First pass--just handle <nowiki> sections, pass the rest off
83 # to internalParse() which does all the real work.
85 # Returns a ParserOutput
87 function parse( $text, &$title, $options, $linestart = true, $clearState = true )
90 $fname = "Parser::parse";
91 wfProfileIn( $fname );
97 $this->mOptions
= $options;
98 $this->mTitle
=& $title;
99 $this->mOutputType
= OT_HTML
;
102 $text = $this->strip( $text, $this->mStripState
);
103 $text = $this->internalParse( $text, $linestart );
104 $text = $this->unstrip( $text, $this->mStripState
);
105 # Clean up special characters, only run once, next-to-last before doBlockLevels
108 "/<hr *>/i" => '<hr/>',
109 "/<br *>/i" => '<br/>',
110 "/<center *>/i"=>'<div class="center">',
111 "/<\\/center *>/i" => '</div>',
112 # Clean up spare ampersands; note that we probably ought to be
113 # more careful about named entities.
114 '/&(?!:amp;|#[Xx][0-9A-fa-f]+;|#[0-9]+;|[a-zA-Z0-9]+;)/' => '&'
116 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
119 "/<center *>/i"=>'<div class="center">',
120 "/<\\/center *>/i" => '</div>'
122 $text = preg_replace( array_keys($fixtags), array_values($fixtags), $text );
125 $text = $this->doBlockLevels( $text, $linestart );
127 $text = $this->tidy($text);
129 $this->mOutput
->setText( $text );
130 wfProfileOut( $fname );
131 return $this->mOutput
;
134 /* static */ function getRandomString()
136 return dechex(mt_rand(0, 0x7fffffff)) . dechex(mt_rand(0, 0x7fffffff));
139 # Replaces all occurrences of <$tag>content</$tag> in the text
140 # with a random marker and returns the new text. the output parameter
141 # $content will be an associative array filled with data on the form
142 # $unique_marker => content.
144 # If $content is already set, the additional entries will be appended
146 # If $tag is set to STRIP_COMMENTS, the function will extract
147 # <!-- HTML comments -->
149 /* static */ function extractTags($tag, $text, &$content, $uniq_prefix = ""){
150 $rnd = $uniq_prefix . '-' . $tag . Parser
::getRandomString();
157 while ( "" != $text ) {
158 if($tag==STRIP_COMMENTS
) {
159 $p = preg_split( "/<!--/i", $text, 2 );
161 $p = preg_split( "/<\\s*$tag\\s*>/i", $text, 2 );
164 if ( ( count( $p ) < 2 ) ||
( "" == $p[1] ) ) {
167 if($tag==STRIP_COMMENTS
) {
168 $q = preg_split( "/-->/i", $p[1], 2 );
170 $q = preg_split( "/<\\/\\s*$tag\\s*>/i", $p[1], 2 );
172 $marker = $rnd . sprintf("%08X", $n++
);
173 $content[$marker] = $q[0];
174 $stripped .= $marker;
181 # Strips and renders <nowiki>, <pre>, <math>, <hiero>
182 # If $render is set, performs necessary rendering operations on plugins
183 # Returns the text, and fills an array with data needed in unstrip()
184 # If the $state is already a valid strip state, it adds to the state
186 # When $stripcomments is set, HTML comments <!-- like this -->
187 # will be stripped in addition to other tags. This is important
188 # for section editing, where these comments cause confusion when
189 # counting the sections in the wikisource
190 function strip( $text, &$state, $stripcomments = false )
192 $render = ($this->mOutputType
== OT_HTML
);
193 $nowiki_content = array();
194 $hiero_content = array();
195 $math_content = array();
196 $pre_content = array();
197 $comment_content = array();
199 # Replace any instances of the placeholders
200 $uniq_prefix = UNIQ_PREFIX
;
201 #$text = str_replace( $uniq_prefix, wfHtmlEscapeFirst( $uniq_prefix ), $text );
203 $text = Parser
::extractTags("nowiki", $text, $nowiki_content, $uniq_prefix);
204 foreach( $nowiki_content as $marker => $content ){
206 $nowiki_content[$marker] = wfEscapeHTMLTagsOnly( $content );
208 $nowiki_content[$marker] = "<nowiki>$content</nowiki>";
212 $text = Parser
::extractTags("hiero", $text, $hiero_content, $uniq_prefix);
213 foreach( $hiero_content as $marker => $content ){
214 if( $render && $GLOBALS['wgUseWikiHiero']){
215 $hiero_content[$marker] = WikiHiero( $content, WH_MODE_HTML
);
217 $hiero_content[$marker] = "<hiero>$content</hiero>";
221 $text = Parser
::extractTags("math", $text, $math_content, $uniq_prefix);
222 foreach( $math_content as $marker => $content ){
224 if( $this->mOptions
->getUseTeX() ) {
225 $math_content[$marker] = renderMath( $content );
227 $math_content[$marker] = "<math>$content<math>";
230 $math_content[$marker] = "<math>$content</math>";
234 $text = Parser
::extractTags("pre", $text, $pre_content, $uniq_prefix);
235 foreach( $pre_content as $marker => $content ){
237 $pre_content[$marker] = "<pre>" . wfEscapeHTMLTagsOnly( $content ) . "</pre>";
239 $pre_content[$marker] = "<pre>$content</pre>";
243 $text = Parser
::extractTags(STRIP_COMMENTS
, $text, $comment_content, $uniq_prefix);
244 foreach( $comment_content as $marker => $content ){
245 $comment_content[$marker] = "<!--$content-->";
249 # Merge state with the pre-existing state, if there is one
251 $state['nowiki'] = $state['nowiki'] +
$nowiki_content;
252 $state['hiero'] = $state['hiero'] +
$hiero_content;
253 $state['math'] = $state['math'] +
$math_content;
254 $state['pre'] = $state['pre'] +
$pre_content;
255 $state['comment'] = $state['comment'] +
$comment_content;
258 'nowiki' => $nowiki_content,
259 'hiero' => $hiero_content,
260 'math' => $math_content,
261 'pre' => $pre_content,
262 'comment' => $comment_content
268 function unstrip( $text, &$state )
270 # Must expand in reverse order, otherwise nested tags will be corrupted
271 $contentDict = end( $state );
272 for ( $contentDict = end( $state ); $contentDict !== false; $contentDict = prev( $state ) ) {
273 for ( $content = end( $contentDict ); $content !== false; $content = prev( $contentDict ) ) {
274 $text = str_replace( key( $contentDict ), $content, $text );
281 # Add an item to the strip state
282 # Returns the unique tag which must be inserted into the stripped text
283 # The tag will be replaced with the original text in unstrip()
285 function insertStripItem( $text, &$state )
287 $rnd = UNIQ_PREFIX
. '-item' . Parser
::getRandomString();
296 $state['item'][$rnd] = $text;
300 # This method generates the list of subcategories and pages for a category
301 function categoryMagic ()
303 global $wgLang , $wgUser ;
304 if ( !$this->mOptions
->getUseCategoryMagic() ) return ; # Doesn't use categories at all
306 $cns = Namespace::getCategory() ;
307 if ( $this->mTitle
->getNamespace() != $cns ) return "" ; # This ain't a category page
309 $r = "<br style=\"clear:both;\"/>\n";
312 $sk =& $wgUser->getSkin() ;
314 $articles = array() ;
315 $children = array() ;
317 $id = $this->mTitle
->getArticleID() ;
319 # For existing categories
321 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM cur,links WHERE l_to={$id} AND l_from=cur_id";
322 $res = wfQuery ( $sql, DB_READ
) ;
323 while ( $x = wfFetchObject ( $res ) ) $data[] = $x ;
325 # For non-existing categories
326 $t = wfStrencode( $this->mTitle
->getPrefixedDBKey() );
327 $sql = "SELECT DISTINCT cur_title,cur_namespace FROM cur,brokenlinks WHERE bl_to='$t' AND bl_from=cur_id" ;
328 $res = wfQuery ( $sql, DB_READ
) ;
329 while ( $x = wfFetchObject ( $res ) ) $data[] = $x ;
332 # For all pages that link to this category
333 foreach ( $data AS $x )
335 $t = $wgLang->getNsText ( $x->cur_namespace
) ;
336 if ( $t != "" ) $t .= ":" ;
337 $t .= $x->cur_title
;
339 if ( $x->cur_namespace
== $cns ) {
340 array_push ( $children , $sk->makeLink ( $t ) ) ; # Subcategory
342 array_push ( $articles , $sk->makeLink ( $t ) ) ; # Page in this category
345 wfFreeResult ( $res ) ;
347 # Showing subcategories
348 if ( count ( $children ) > 0 )
350 asort ( $children ) ;
351 $r .= "<h2>".wfMsg("subcategories")."</h2>\n" ;
352 $r .= implode ( ", " , $children ) ;
355 # Showing pages in this category
356 if ( count ( $articles ) > 0 )
358 $ti = $this->mTitle
->getText() ;
359 asort ( $articles ) ;
360 $h = wfMsg( "category_header", $ti );
361 $r .= "<h2>{$h}</h2>\n" ;
362 $r .= implode ( ", " , $articles ) ;
369 function getHTMLattrs ()
371 $htmlattrs = array( # Allowed attributes--no scripting, etc.
372 "title", "align", "lang", "dir", "width", "height",
373 "bgcolor", "clear", /* BR */ "noshade", /* HR */
374 "cite", /* BLOCKQUOTE, Q */ "size", "face", "color",
375 /* FONT */ "type", "start", "value", "compact",
376 /* For various lists, mostly deprecated but safe */
377 "summary", "width", "border", "frame", "rules",
378 "cellspacing", "cellpadding", "valign", "char",
379 "charoff", "colgroup", "col", "span", "abbr", "axis",
380 "headers", "scope", "rowspan", "colspan", /* Tables */
381 "id", "class", "name", "style" /* For CSS */
386 function fixTagAttributes ( $t )
388 if ( trim ( $t ) == "" ) return "" ; # Saves runtime ;-)
389 $htmlattrs = $this->getHTMLattrs() ;
391 # Strip non-approved attributes from the tag
393 "/(\\w+)(\\s*=\\s*([^\\s\">]+|\"[^\">]*\"))?/e",
394 "(in_array(strtolower(\"\$1\"),\$htmlattrs)?(\"\$1\".((\"x\$3\" != \"x\")?\"=\$3\":'')):'')",
396 # Strip javascript "expression" from stylesheets. Brute force approach:
397 # If anythin offensive is found, all attributes of the HTML tag are dropped
400 "/style\\s*=.*(expression|tps*:\/\/|url\\s*\().*/is",
401 wfMungeToUtf8( $t ) ) )
409 /* interface with html tidy, used if $wgUseTidy = true */
410 function tidy ( $text ) {
411 global $wgTidyConf, $wgTidyBin, $wgTidyOpts;
412 global $wgInputEncoding, $wgOutputEncoding;
414 switch(strtoupper($wgOutputEncoding)) {
416 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)?
' -latin1':' -raw';
419 $wgTidyOpts .= ($wgInputEncoding == $wgOutputEncoding)?
' -utf8':' -raw';
422 $wgTidyOpts .= ' -raw';
425 $text = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"'.
426 ' "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html>'.
427 '<head><title>test</title></head><body>'.$text.'</body></html>';
428 $descriptorspec = array(
429 0 => array("pipe", "r"),
430 1 => array("pipe", "w"),
431 2 => array("file", "/dev/null", "a")
433 $process = proc_open("$wgTidyBin -config $wgTidyConf $wgTidyOpts", $descriptorspec, $pipes);
434 if (is_resource($process)) {
435 fwrite($pipes[0], $text);
437 while (!feof($pipes[1])) {
438 $cleansource .= fgets($pipes[1], 1024);
441 $return_value = proc_close($process);
443 if( $cleansource == '' && $text != '') {
444 return '<h2>'.wfMsg('seriousxhtmlerrors').'</h2><pre>'.htmlspecialchars($text).'</pre>';
450 function doTableStuff ( $t )
452 $t = explode ( "\n" , $t ) ;
453 $td = array () ; # Is currently a td tag open?
454 $ltd = array () ; # Was it TD or TH?
455 $tr = array () ; # Is currently a tr tag open?
456 $ltr = array () ; # tr attributes
457 foreach ( $t AS $k => $x )
460 $fc = substr ( $x , 0 , 1 ) ;
461 if ( "{|" == substr ( $x , 0 , 2 ) )
463 $t[$k] = "\n<table " . $this->fixTagAttributes ( substr ( $x , 3 ) ) . ">" ;
464 array_push ( $td , false ) ;
465 array_push ( $ltd , "" ) ;
466 array_push ( $tr , false ) ;
467 array_push ( $ltr , "" ) ;
469 else if ( count ( $td ) == 0 ) { } # Don't do any of the following
470 else if ( "|}" == substr ( $x , 0 , 2 ) )
473 $l = array_pop ( $ltd ) ;
474 if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
475 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
479 /* else if ( "|_" == substr ( $x , 0 , 2 ) ) # Caption
481 $z = trim ( substr ( $x , 2 ) ) ;
482 $t[$k] = "<caption>{$z}</caption>\n" ;
484 else if ( "|-" == substr ( $x , 0 , 2 ) ) # Allows for |---------------
486 $x = substr ( $x , 1 ) ;
487 while ( $x != "" && substr ( $x , 0 , 1 ) == '-' ) $x = substr ( $x , 1 ) ;
489 $l = array_pop ( $ltd ) ;
490 if ( array_pop ( $tr ) ) $z = "</tr>" . $z ;
491 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
494 array_push ( $tr , false ) ;
495 array_push ( $td , false ) ;
496 array_push ( $ltd , "" ) ;
497 array_push ( $ltr , $this->fixTagAttributes ( $x ) ) ;
499 else if ( "|" == $fc ||
"!" == $fc ||
"|+" == substr ( $x , 0 , 2 ) ) # Caption
501 if ( "|+" == substr ( $x , 0 , 2 ) )
504 $x = substr ( $x , 1 ) ;
506 $after = substr ( $x , 1 ) ;
507 if ( $fc == "!" ) $after = str_replace ( "!!" , "||" , $after ) ;
508 $after = explode ( "||" , $after ) ;
510 foreach ( $after AS $theline )
515 $tra = array_pop ( $ltr ) ;
516 if ( !array_pop ( $tr ) ) $z = "<tr {$tra}>\n" ;
517 array_push ( $tr , true ) ;
518 array_push ( $ltr , "" ) ;
521 $l = array_pop ( $ltd ) ;
522 if ( array_pop ( $td ) ) $z = "</{$l}>" . $z ;
523 if ( $fc == "|" ) $l = "td" ;
524 else if ( $fc == "!" ) $l = "th" ;
525 else if ( $fc == "+" ) $l = "caption" ;
527 array_push ( $ltd , $l ) ;
528 $y = explode ( "|" , $theline , 2 ) ;
529 if ( count ( $y ) == 1 ) $y = "{$z}<{$l}>{$y[0]}" ;
530 else $y = $y = "{$z}<{$l} ".$this->fixTagAttributes($y[0]).">{$y[1]}" ;
532 array_push ( $td , true ) ;
537 # Closing open td, tr && table
538 while ( count ( $td ) > 0 )
540 if ( array_pop ( $td ) ) $t[] = "</td>" ;
541 if ( array_pop ( $tr ) ) $t[] = "</tr>" ;
545 $t = implode ( "\n" , $t ) ;
546 # $t = $this->removeHTMLtags( $t );
550 function internalParse( $text, $linestart, $args = array() )
552 $fname = "Parser::internalParse";
553 wfProfileIn( $fname );
555 $text = $this->removeHTMLtags( $text );
556 $text = $this->replaceVariables( $text, $args );
558 # $text = preg_replace( "/(^|\n)-----*/", "\\1<hr>", $text );
560 $text = $this->doHeadings( $text );
561 if($this->mOptions
->getUseDynamicDates()) {
562 global $wgDateFormatter;
563 $text = $wgDateFormatter->reformat( $this->mOptions
->getDateFormat(), $text );
565 $text = $this->replaceExternalLinks( $text );
566 $text = $this->doTokenizedParser ( $text );
567 $text = $this->doTableStuff ( $text ) ;
568 $text = $this->formatHeadings( $text );
569 $sk =& $this->mOptions
->getSkin();
570 $text = $sk->transformContent( $text );
572 if ( !isset ( $this->categoryMagicDone
) ) {
573 $text .= $this->categoryMagic () ;
574 $this->categoryMagicDone
= true ;
577 wfProfileOut( $fname );
582 /* private */ function doHeadings( $text )
584 for ( $i = 6; $i >= 1; --$i ) {
585 $h = substr( "======", 0, $i );
586 $text = preg_replace( "/^{$h}(.+){$h}(\\s|$)/m",
587 "<h{$i}>\\1</h{$i}>\\2", $text );
592 # Note: we have to do external links before the internal ones,
593 # and otherwise take great care in the order of things here, so
594 # that we don't end up interpreting some URLs twice.
596 /* private */ function replaceExternalLinks( $text )
598 $fname = "Parser::replaceExternalLinks";
599 wfProfileIn( $fname );
600 $text = $this->subReplaceExternalLinks( $text, "http", true );
601 $text = $this->subReplaceExternalLinks( $text, "https", true );
602 $text = $this->subReplaceExternalLinks( $text, "ftp", false );
603 $text = $this->subReplaceExternalLinks( $text, "irc", false );
604 $text = $this->subReplaceExternalLinks( $text, "gopher", false );
605 $text = $this->subReplaceExternalLinks( $text, "news", false );
606 $text = $this->subReplaceExternalLinks( $text, "mailto", false );
607 wfProfileOut( $fname );
611 /* private */ function subReplaceExternalLinks( $s, $protocol, $autonumber )
613 $unique = "4jzAfzB8hNvf4sqyO9Edd8pSmk9rE2in0Tgw3";
614 $uc = "A-Za-z0-9_\\/~%\\-+&*#?!=()@\\x80-\\xFF";
616 # this is the list of separators that should be ignored if they
617 # are the last character of an URL but that should be included
618 # if they occur within the URL, e.g. "go to www.foo.com, where .."
619 # in this case, the last comma should not become part of the URL,
620 # but in "www.foo.com/123,2342,32.htm" it should.
622 $fnc = "A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF";
623 $images = "gif|png|jpg|jpeg";
625 # PLEASE NOTE: The curly braces { } are not part of the regex,
626 # they are interpreted as part of the string (used to tell PHP
627 # that the content of the string should be inserted there).
628 $e1 = "/(^|[^\\[])({$protocol}:)([{$uc}{$sep}]+)\\/([{$fnc}]+)\\." .
629 "((?i){$images})([^{$uc}]|$)/";
631 $e2 = "/(^|[^\\[])({$protocol}:)(([".$uc."]|[".$sep."][".$uc."])+)([^". $uc . $sep. "]|[".$sep."]|$)/";
632 $sk =& $this->mOptions
->getSkin();
634 if ( $autonumber and $this->mOptions
->getAllowExternalImages() ) { # Use img tags only for HTTP urls
635 $s = preg_replace( $e1, "\\1" . $sk->makeImage( "{$unique}:\\3" .
636 "/\\4.\\5", "\\4.\\5" ) . "\\6", $s );
638 $s = preg_replace( $e2, "\\1" . "<a href=\"{$unique}:\\3\"" .
639 $sk->getExternalLinkAttributes( "{$unique}:\\3", wfEscapeHTML(
640 "{$unique}:\\3" ) ) . ">" . wfEscapeHTML( "{$unique}:\\3" ) .
642 $s = str_replace( $unique, $protocol, $s );
644 $a = explode( "[{$protocol}:", " " . $s );
645 $s = array_shift( $a );
646 $s = substr( $s, 1 );
648 $e1 = "/^([{$uc}"."{$sep}]+)](.*)\$/sD";
649 $e2 = "/^([{$uc}"."{$sep}]+)\\s+([^\\]]+)](.*)\$/sD";
651 foreach ( $a as $line ) {
652 if ( preg_match( $e1, $line, $m ) ) {
653 $link = "{$protocol}:{$m[1]}";
655 if ( $autonumber ) { $text = "[" . ++
$this->mAutonumber
. "]"; }
656 else { $text = wfEscapeHTML( $link ); }
657 } else if ( preg_match( $e2, $line, $m ) ) {
658 $link = "{$protocol}:{$m[1]}";
662 $s .= "[{$protocol}:" . $line;
665 if( $link == $text ||
preg_match( "!$protocol://" . preg_quote( $text, "/" ) . "/?$!", $link ) ) {
668 # Expand the URL for printable version
669 $paren = "<span class='urlexpansion'> (<i>" . htmlspecialchars ( $link ) . "</i>)</span>";
671 $la = $sk->getExternalLinkAttributes( $link, $text );
672 $s .= "<a href='{$link}'{$la}>{$text}</a>{$paren}{$trail}";
678 /* private */ function handle3Quotes( &$state, $token )
680 if ( $state["strong"] !== false ) {
681 if ( $state["em"] !== false && $state["em"] > $state["strong"] )
683 # ''' lala ''lala '''
684 $s = "</em></strong><em>";
688 $state["strong"] = FALSE;
691 $state["strong"] = isset($token["pos"]) ?
$token["pos"] : true;
696 /* private */ function handle2Quotes( &$state, $token )
698 if ( $state["em"] !== false ) {
699 if ( $state["strong"] !== false && $state["strong"] > $state["em"] )
701 # ''lala'''lala'' ....'''
702 $s = "</strong></em><strong>";
706 $state["em"] = FALSE;
709 $state["em"] = isset($token["pos"]) ?
$token["pos"] : true;
715 /* private */ function handle5Quotes( &$state, $token )
718 if ( $state["em"] !== false && $state["strong"] !== false ) {
719 if ( $state["em"] < $state["strong"] ) {
720 $s .= "</strong></em>";
722 $s .= "</em></strong>";
724 $state["strong"] = $state["em"] = FALSE;
725 } elseif ( $state["em"] !== false ) {
726 $s .= "</em><strong>";
727 $state["em"] = FALSE;
728 $state["strong"] = $token["pos"];
729 } elseif ( $state["strong"] !== false ) {
730 $s .= "</strong><em>";
731 $state["strong"] = FALSE;
732 $state["em"] = $token["pos"];
733 } else { # not $em and not $strong
734 $s .= "<strong><em>";
735 $state["strong"] = $state["em"] = isset($token["pos"]) ?
$token["pos"] : true;
740 /* private */ function doTokenizedParser( $str )
742 global $wgLang; # for language specific parser hook
743 global $wgUploadDirectory, $wgUseTimeline;
745 $tokenizer=Tokenizer
::newFromString( $str );
746 $tokenStack = array();
749 $state["em"] = FALSE;
750 $state["strong"] = FALSE;
754 # The tokenizer splits the text into tokens and returns them one by one.
755 # Every call to the tokenizer returns a new token.
756 while ( $token = $tokenizer->nextToken() )
758 switch ( $token["type"] )
761 # simple text with no further markup
762 $txt = $token["text"];
765 # Text that contains blanks that have to be converted to
766 # non-breakable spaces for French.
767 # U+202F NARROW NO-BREAK SPACE might be a better choice, but
768 # browser support for Unicode spacing is poor.
769 $txt = str_replace( " ", " ", $token["text"] );
772 # remember the tag opened with 3 [
776 # FIXME : Treat orphaned open tags (stack not empty when text is over)
778 array_push( $tokenStack, $token );
785 # get text from stack, glue it together, and call the code to handle a
788 if ( count( $tokenStack ) == 0 )
790 # stack empty. Found a ]] without an opening [[
794 $lastToken = array_pop( $tokenStack );
795 while ( !(($lastToken["type"] == "[[[") or ($lastToken["type"] == "[[")) )
797 if( !empty( $lastToken["text"] ) ) {
798 $linkText = $lastToken["text"] . $linkText;
800 $lastToken = array_pop( $tokenStack );
803 $txt = $linkText ."]]";
805 if( isset( $lastToken["text"] ) ) {
806 $prefix = $lastToken["text"];
810 $nextToken = $tokenizer->previewToken();
811 if ( $nextToken["type"] == "text" )
813 # Preview just looks at it. Now we have to fetch it.
814 $nextToken = $tokenizer->nextToken();
815 $txt .= $nextToken["text"];
817 $txt = $this->handleInternalLink( $this->unstrip($txt,$this->mStripState
), $prefix );
819 # did the tag start with 3 [ ?
821 # show the first as text
827 $tagIsOpen = (count( $tokenStack ) != 0);
833 # This and the three next ones handle quotes
834 $txt = $this->handle3Quotes( $state, $token );
837 $txt = $this->handle2Quotes( $state, $token );
840 $txt = $this->handle5Quotes( $state, $token );
850 $txt = $this->doMagicRFC( $tokenizer );
857 $txt = $this->doMagicISBN( $tokenizer );
861 if ( $wgUseTimeline &&
862 "" != ( $timelinesrc = $tokenizer->readAllUntil("</timeline>") ) )
864 $txt = renderTimeline( $timelinesrc );
870 # Call language specific Hook.
871 $txt = $wgLang->processToken( $token, $tokenStack );
872 if ( NULL == $txt ) {
873 # An unkown token. Highlight.
874 $txt = "<font color=\"#FF0000\"><b>".$token["type"]."</b></font>";
875 $txt .= "<font color=\"#FFFF00\"><b>".$token["text"]."</b></font>";
879 # If we're parsing the interior of a link, don't append the interior to $s,
880 # but push it to the stack so it can be processed when a ]] token is found.
881 if ( $tagIsOpen && $txt != "" ) {
882 $token["type"] = "text";
883 $token["text"] = $txt;
884 array_push( $tokenStack, $token );
889 if ( count( $tokenStack ) != 0 )
891 # still objects on stack. opened [[ tag without closing ]] tag.
893 while ( $lastToken = array_pop( $tokenStack ) )
895 if ( $lastToken["type"] == "text" )
897 $txt = $lastToken["text"] . $txt;
899 $txt = $lastToken["type"] . $txt;
907 /* private */ function handleInternalLink( $line, $prefix )
909 global $wgLang, $wgLinkCache;
910 global $wgNamespacesWithSubpages, $wgLanguageCode;
911 static $fname = "Parser::handleInternalLink" ;
912 wfProfileIn( $fname );
914 wfProfileIn( "$fname-setup" );
916 if ( !$tc ) { $tc = Title
::legalChars() . "#"; }
917 $sk =& $this->mOptions
->getSkin();
919 # Match a link having the form [[namespace:link|alternate]]trail
921 if ( !$e1 ) { $e1 = "/^([{$tc}]+)(?:\\|([^]]+))?]](.*)\$/sD"; }
922 # Match the end of a line for a word that's not followed by whitespace,
923 # e.g. in the case of 'The Arab al[[Razi]]', 'al' will be matched
924 #$e2 = "/^(.*)\\b(\\w+)\$/suD";
925 #$e2 = "/^(.*\\s)(\\S+)\$/suD";
926 static $e2 = '/^(.*\s)([a-zA-Z\x80-\xff]+)$/sD';
929 # Special and Media are pseudo-namespaces; no pages actually exist in them
930 static $image = FALSE;
931 static $special = FALSE;
932 static $media = FALSE;
933 static $category = FALSE;
934 if ( !$image ) { $image = Namespace::getImage(); }
935 if ( !$special ) { $special = Namespace::getSpecial(); }
936 if ( !$media ) { $media = Namespace::getMedia(); }
937 if ( !$category ) { $category = Namespace::getCategory(); }
939 $nottalk = !Namespace::isTalk( $this->mTitle
->getNamespace() );
941 wfProfileOut( "$fname-setup" );
944 if ( preg_match( $e1, $line, $m ) ) { # page with normal text or alt
947 } else { # Invalid form; output directly
948 $s .= $prefix . "[[" . $line ;
954 :Foobar -- override special treatment of prefix (images, language links)
955 /Foobar -- convert to CurrentPage/Foobar
956 /Foobar/ -- convert to CurrentPage/Foobar, strip the initial / from text
958 $c = substr($m[1],0,1);
959 $noforce = ($c != ":");
960 if( $c == "/" ) { # subpage
961 if(substr($m[1],-1,1)=="/") { # / at end means we don't want the slash to be shown
962 $m[1]=substr($m[1],1,strlen($m[1])-2);
965 $noslash=substr($m[1],1);
967 if($wgNamespacesWithSubpages[$this->mTitle
->getNamespace()]) { # subpages allowed here
968 $link = $this->mTitle
->getPrefixedText(). "/" . trim($noslash);
971 } # this might be changed for ugliness reasons
973 $link = $noslash; # no subpage allowed, use standard link
975 } elseif( $noforce ) { # no subpage
978 $link = substr( $m[1], 1 );
983 $nt = Title
::newFromText( $link );
985 $s .= $prefix . "[[" . $line;
988 $ns = $nt->getNamespace();
989 $iw = $nt->getInterWiki();
991 if( $iw && $this->mOptions
->getInterwikiMagic() && $nottalk && $wgLang->getLanguageName( $iw ) ) {
992 array_push( $this->mOutput
->mLanguageLinks
, $nt->getPrefixedText() );
993 $s .= $prefix . $trail ;
994 return (trim($s) == '')?
'': $s;
996 if( $ns == $image ) {
997 $s .= $prefix . $sk->makeImageLinkObj( $nt, $text ) . $trail;
998 $wgLinkCache->addImageLinkObj( $nt );
1001 if ( $ns == $category ) {
1002 $t = $nt->getText() ;
1003 $nnt = Title
::newFromText ( Namespace::getCanonicalName($category).":".$t ) ;
1004 $t = $sk->makeLinkObj( $nnt, $t, "", "" , $prefix );
1005 $this->mOutput
->mCategoryLinks
[] = $t ;
1006 $s .= $prefix . $trail ;
1010 if( ( $nt->getPrefixedText() == $this->mTitle
->getPrefixedText() ) &&
1011 ( strpos( $link, "#" ) == FALSE ) ) {
1012 # Self-links are handled specially; generally de-link and change to bold.
1013 $s .= $prefix . $sk->makeSelfLinkObj( $nt, $text, "", $trail );
1017 if( $ns == $media ) {
1018 $s .= $prefix . $sk->makeMediaLinkObj( $nt, $text ) . $trail;
1019 $wgLinkCache->addImageLinkObj( $nt );
1021 } elseif( $ns == $special ) {
1022 $s .= $prefix . $sk->makeKnownLinkObj( $nt, $text, "", $trail );
1025 $s .= $sk->makeLinkObj( $nt, $text, "", $trail , $prefix );
1027 wfProfileOut( $fname );
1031 # Some functions here used by doBlockLevels()
1033 /* private */ function closeParagraph()
1036 if ( '' != $this->mLastSection
) {
1037 $result = "</" . $this->mLastSection
. ">\n";
1039 $this->mInPre
= false;
1040 $this->mLastSection
= "";
1043 # getCommon() returns the length of the longest common substring
1044 # of both arguments, starting at the beginning of both.
1046 /* private */ function getCommon( $st1, $st2 )
1048 $fl = strlen( $st1 );
1049 $shorter = strlen( $st2 );
1050 if ( $fl < $shorter ) { $shorter = $fl; }
1052 for ( $i = 0; $i < $shorter; ++
$i ) {
1053 if ( $st1{$i} != $st2{$i} ) { break; }
1057 # These next three functions open, continue, and close the list
1058 # element appropriate to the prefix character passed into them.
1060 /* private */ function openList( $char )
1062 $result = $this->closeParagraph();
1064 if ( "*" == $char ) { $result .= "<ul><li>"; }
1065 else if ( "#" == $char ) { $result .= "<ol><li>"; }
1066 else if ( ":" == $char ) { $result .= "<dl><dd>"; }
1067 else if ( ";" == $char ) {
1068 $result .= "<dl><dt>";
1069 $this->mDTopen
= true;
1071 else { $result = "<!-- ERR 1 -->"; }
1076 /* private */ function nextItem( $char )
1078 if ( "*" == $char ||
"#" == $char ) { return "</li><li>"; }
1079 else if ( ":" == $char ||
";" == $char ) {
1081 if ( $this->mDTopen
) { $close = "</dt>"; }
1082 if ( ";" == $char ) {
1083 $this->mDTopen
= true;
1084 return $close . "<dt>";
1086 $this->mDTopen
= false;
1087 return $close . "<dd>";
1090 return "<!-- ERR 2 -->";
1093 /* private */function closeList( $char )
1095 if ( "*" == $char ) { $text = "</li></ul>"; }
1096 else if ( "#" == $char ) { $text = "</li></ol>"; }
1097 else if ( ":" == $char ) {
1098 if ( $this->mDTopen
) {
1099 $this->mDTopen
= false;
1100 $text = "</dt></dl>";
1102 $text = "</dd></dl>";
1105 else { return "<!-- ERR 3 -->"; }
1109 /* private */ function doBlockLevels( $text, $linestart ) {
1110 $fname = "Parser::doBlockLevels";
1111 wfProfileIn( $fname );
1113 # Parsing through the text line by line. The main thing
1114 # happening here is handling of block-level elements p, pre,
1115 # and making lists from lines starting with * # : etc.
1117 $textLines = explode( "\n", $text );
1119 $lastPrefix = $output = $lastLine = '';
1120 $this->mDTopen
= $inBlockElem = false;
1122 $paragraphStack = false;
1124 if ( !$linestart ) {
1125 $output .= array_shift( $textLines );
1127 foreach ( $textLines as $oLine ) {
1128 $lastPrefixLength = strlen( $lastPrefix );
1129 $preCloseMatch = preg_match("/<\\/pre/i", $oLine );
1130 $preOpenMatch = preg_match("/<pre/i", $oLine );
1131 if (!$this->mInPre
) {
1132 $this->mInPre
= !empty($preOpenMatch);
1134 if ( !$this->mInPre
) {
1135 # Multiple prefixes may abut each other for nested lists.
1136 $prefixLength = strspn( $oLine, "*#:;" );
1137 $pref = substr( $oLine, 0, $prefixLength );
1140 $pref2 = str_replace( ";", ":", $pref );
1141 $t = substr( $oLine, $prefixLength );
1143 # Don't interpret any other prefixes in preformatted text
1145 $pref = $pref2 = '';
1150 if( $prefixLength && 0 == strcmp( $lastPrefix, $pref2 ) ) {
1151 # Same as the last item, so no need to deal with nesting or opening stuff
1152 $output .= $this->nextItem( substr( $pref, -1 ) );
1153 $paragraphStack = false;
1155 if ( ";" == substr( $pref, -1 ) ) {
1156 # The one nasty exception: definition lists work like this:
1157 # ; title : definition text
1158 # So we check for : in the remainder text to split up the
1159 # title and definition, without b0rking links.
1160 # FIXME: This is not foolproof. Something better in Tokenizer might help.
1161 if( preg_match( '/^(.*?(?:\s| )):(.*)$/', $t, $match ) ) {
1163 $output .= $term . $this->nextItem( ":" );
1167 } elseif( $prefixLength ||
$lastPrefixLength ) {
1168 # Either open or close a level...
1169 $commonPrefixLength = $this->getCommon( $pref, $lastPrefix );
1170 $paragraphStack = false;
1172 while( $commonPrefixLength < $lastPrefixLength ) {
1173 $output .= $this->closeList( $lastPrefix{$lastPrefixLength-1} );
1174 --$lastPrefixLength;
1176 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
1177 $output .= $this->nextItem( $pref{$commonPrefixLength-1} );
1179 while ( $prefixLength > $commonPrefixLength ) {
1180 $char = substr( $pref, $commonPrefixLength, 1 );
1181 $output .= $this->openList( $char );
1183 if ( ";" == $char ) {
1184 # FIXME: This is dupe of code above
1185 if( preg_match( '/^(.*?(?:\s| )):(.*)$/', $t, $match ) ) {
1187 $output .= $term . $this->nextItem( ":" );
1191 ++
$commonPrefixLength;
1193 $lastPrefix = $pref2;
1195 if( 0 == $prefixLength ) {
1196 # No prefix (not in list)--go to paragraph mode
1197 $uniq_prefix = UNIQ_PREFIX
;
1198 // XXX: use a stack for nestable elements like span, table and div
1199 $openmatch = preg_match("/(<table|<blockquote|<h1|<h2|<h3|<h4|<h5|<h6|<div|<pre|<tr|<td|<p|<ul|<li)/i", $t );
1200 $closematch = preg_match(
1201 "/(<\\/table|<\\/blockquote|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|".
1202 "<\\/div|<hr|<\\/td|<\\/pre|<\\/p|".$uniq_prefix."-pre|<\\/li|<\\/ul)/i", $t );
1203 if ( $openmatch or $closematch ) {
1204 $paragraphStack = false;
1205 $output .= $this->closeParagraph();
1206 if($preOpenMatch and !$preCloseMatch) {
1207 $this->mInPre
= true;
1209 if ( $closematch ) {
1210 $inBlockElem = false;
1212 $inBlockElem = true;
1214 } else if ( !$inBlockElem ) {
1215 if ( " " == $t{0} ) {
1217 if ($this->mLastSection
!= 'pre') {
1218 $paragraphStack = false;
1219 $output .= $this->closeParagraph().'<pre>';
1220 $this->mLastSection
= 'pre';
1224 if ( '' == trim($t) ) {
1225 if ( $paragraphStack ) {
1226 $output .= $paragraphStack.'<br/>';
1227 $paragraphStack = false;
1228 $this->mLastSection
= 'p';
1230 if ($this->mLastSection
!= 'p' ) {
1231 $output .= $this->closeParagraph();
1232 $this->mLastSection
= '';
1233 $paragraphStack = "<p>";
1235 $paragraphStack = '</p><p>';
1239 if ( $paragraphStack ) {
1240 $output .= $paragraphStack;
1241 $paragraphStack = false;
1242 $this->mLastSection
= 'p';
1243 } else if ($this->mLastSection
!= 'p') {
1244 $output .= $this->closeParagraph().'<p>';
1245 $this->mLastSection
= 'p';
1251 if ($paragraphStack === false) {
1255 while ( $prefixLength ) {
1256 $output .= $this->closeList( $pref2{$prefixLength-1} );
1259 if ( "" != $this->mLastSection
) {
1260 $output .= "</" . $this->mLastSection
. ">";
1261 $this->mLastSection
= "";
1264 wfProfileOut( $fname );
1268 function getVariableValue( $index ) {
1269 global $wgLang, $wgSitename, $wgServer;
1272 case MAG_CURRENTMONTH
:
1274 case MAG_CURRENTMONTHNAME
:
1275 return $wgLang->getMonthName( date("n") );
1276 case MAG_CURRENTMONTHNAMEGEN
:
1277 return $wgLang->getMonthNameGen( date("n") );
1278 case MAG_CURRENTDAY
:
1281 return $this->mTitle
->getText();
1283 # return Namespace::getCanonicalName($this->mTitle->getNamespace());
1284 return $wgLang->getNsText($this->mTitle
->getNamespace()); // Patch by Dori
1285 case MAG_CURRENTDAYNAME
:
1286 return $wgLang->getWeekdayName( date("w")+
1 );
1287 case MAG_CURRENTYEAR
:
1289 case MAG_CURRENTTIME
:
1290 return $wgLang->time( wfTimestampNow(), false );
1291 case MAG_NUMBEROFARTICLES
:
1292 return wfNumberOfArticles();
1302 function initialiseVariables()
1304 global $wgVariableIDs;
1305 $this->mVariables
= array();
1306 foreach ( $wgVariableIDs as $id ) {
1307 $mw =& MagicWord
::get( $id );
1308 $mw->addToArray( $this->mVariables
, $this->getVariableValue( $id ) );
1312 /* private */ function replaceVariables( $text, $args = array() )
1314 global $wgLang, $wgScript, $wgArticlePath;
1316 $fname = "Parser::replaceVariables";
1317 wfProfileIn( $fname );
1320 if ( !$this->mVariables
) {
1321 $this->initialiseVariables();
1323 $titleChars = Title
::legalChars();
1324 $regex = "/(\\n?){{([$titleChars]*?)(\\|.*?|)}}/s";
1326 # This function is called recursively. To keep track of arguments we need a stack:
1327 array_push( $this->mArgStack
, $args );
1329 # PHP global rebinding syntax is a bit weird, need to use the GLOBALS array
1330 $GLOBALS['wgCurParser'] =& $this;
1331 $text = preg_replace_callback( $regex, "wfBraceSubstitution", $text );
1333 array_pop( $this->mArgStack
);
1338 function braceSubstitution( $matches )
1340 global $wgLinkCache, $wgLang;
1341 $fname = "Parser::braceSubstitution";
1346 # $newline is an optional newline character before the braces
1347 # $part1 is the bit before the first |, and must contain only title characters
1348 # $args is a list of arguments, starting from index 0, not including $part1
1350 $newline = $matches[1];
1351 $part1 = $matches[2];
1352 # If the third subpattern matched anything, it will start with |
1353 if ( $matches[3] !== "" ) {
1354 $args = explode( "|", substr( $matches[3], 1 ) );
1358 $argc = count( $args );
1361 $mwSubst =& MagicWord
::get( MAG_SUBST
);
1362 if ( $mwSubst->matchStartAndRemove( $part1 ) ) {
1363 if ( $this->mOutputType
!= OT_WIKI
) {
1364 # Invalid SUBST not replaced at PST time
1365 # Return without further processing
1366 $text = $matches[0];
1369 } elseif ( $this->mOutputType
== OT_WIKI
) {
1370 # SUBST not found in PST pass, do nothing
1371 $text = $matches[0];
1375 # MSG, MSGNW and INT
1378 $mwMsgnw =& MagicWord
::get( MAG_MSGNW
);
1379 if ( $mwMsgnw->matchStartAndRemove( $part1 ) ) {
1382 # Remove obsolete MSG:
1383 $mwMsg =& MagicWord
::get( MAG_MSG
);
1384 $mwMsg->matchStartAndRemove( $part1 );
1387 # Check if it is an internal message
1388 $mwInt =& MagicWord
::get( MAG_INT
);
1389 if ( $mwInt->matchStartAndRemove( $part1 ) ) {
1390 if ( $this->incrementIncludeCount( "int:$part1" ) ) {
1391 $text = wfMsgReal( $part1, $args, true );
1399 # Check for NS: (namespace expansion)
1400 $mwNs = MagicWord
::get( MAG_NS
);
1401 if ( $mwNs->matchStartAndRemove( $part1 ) ) {
1402 if ( intval( $part1 ) ) {
1403 $text = $wgLang->getNsText( intval( $part1 ) );
1406 $index = Namespace::getCanonicalIndex( strtolower( $part1 ) );
1407 if ( !is_null( $index ) ) {
1408 $text = $wgLang->getNsText( $index );
1415 # LOCALURL and LOCALURLE
1417 $mwLocal = MagicWord
::get( MAG_LOCALURL
);
1418 $mwLocalE = MagicWord
::get( MAG_LOCALURLE
);
1420 if ( $mwLocal->matchStartAndRemove( $part1 ) ) {
1421 $func = 'getLocalURL';
1422 } elseif ( $mwLocalE->matchStartAndRemove( $part1 ) ) {
1423 $func = 'escapeLocalURL';
1428 if ( $func !== '' ) {
1429 $title = Title
::newFromText( $part1 );
1430 if ( !is_null( $title ) ) {
1432 $text = $title->$func( $args[0] );
1434 $text = $title->$func();
1441 # Internal variables
1442 if ( !$found && array_key_exists( $part1, $this->mVariables
) ) {
1443 $text = $this->mVariables
[$part1];
1445 $this->mOutput
->mContainsOldMagic
= true;
1448 # Arguments input from the caller
1449 $inputArgs = end( $this->mArgStack
);
1450 if ( !$found && array_key_exists( $part1, $inputArgs ) ) {
1451 $text = $inputArgs[$part1];
1455 # Load from database
1457 $title = Title
::newFromText( $part1, NS_TEMPLATE
);
1458 if ( !is_null( $title ) && !$title->isExternal() ) {
1459 # Check for excessive inclusion
1460 $dbk = $title->getPrefixedDBkey();
1461 if ( $this->incrementIncludeCount( $dbk ) ) {
1462 $article = new Article( $title );
1463 $articleContent = $article->getContentWithoutUsingSoManyDamnGlobals();
1464 if ( $articleContent !== false ) {
1466 $text = $articleContent;
1471 # If the title is valid but undisplayable, make a link to it
1472 if ( $this->mOutputType
== OT_HTML
&& !$found ) {
1473 $text = "[[" . $title->getPrefixedText() . "]]";
1479 # Recursive parsing, escaping and link table handling
1480 # Only for HTML output
1481 if ( $nowiki && $found && $this->mOutputType
== OT_HTML
) {
1482 $text = wfEscapeWikiText( $text );
1483 } elseif ( $this->mOutputType
== OT_HTML
&& $found ) {
1484 # Clean up argument array
1485 $assocArgs = array();
1487 foreach( $args as $arg ) {
1488 $eqpos = strpos( $arg, "=" );
1489 if ( $eqpos === false ) {
1490 $assocArgs[$index++
] = $arg;
1492 $name = trim( substr( $arg, 0, $eqpos ) );
1493 $value = trim( substr( $arg, $eqpos+
1 ) );
1494 if ( $value === false ) {
1497 if ( $name !== false ) {
1498 $assocArgs[$name] = $value;
1503 # Do not enter included links in link table
1504 if ( !is_null( $title ) ) {
1505 $wgLinkCache->suspend();
1508 # Run full parser on the included text
1509 $text = $this->strip( $text, $this->mStripState
);
1510 $text = $this->internalParse( $text, (bool)$newline, $assocArgs );
1511 if(!empty($newline)) $text = "\n".$text;
1513 # Add the result to the strip state for re-inclusion after
1514 # the rest of the processing
1515 $text = $this->insertStripItem( $text, $this->mStripState
);
1517 # Resume the link cache and register the inclusion as a link
1518 if ( !is_null( $title ) ) {
1519 $wgLinkCache->resume();
1520 $wgLinkCache->addLinkObj( $title );
1531 # Returns true if the function is allowed to include this entity
1532 function incrementIncludeCount( $dbk )
1534 if ( !array_key_exists( $dbk, $this->mIncludeCount
) ) {
1535 $this->mIncludeCount
[$dbk] = 0;
1537 if ( ++
$this->mIncludeCount
[$dbk] <= MAX_INCLUDE_REPEAT
) {
1545 # Cleans up HTML, removes dangerous tags and attributes
1546 /* private */ function removeHTMLtags( $text )
1548 global $wgUseTidy, $wgUserHtml;
1549 $fname = "Parser::removeHTMLtags";
1550 wfProfileIn( $fname );
1553 $htmlpairs = array( # Tags that must be closed
1554 "b", "del", "i", "ins", "u", "font", "big", "small", "sub", "sup", "h1",
1555 "h2", "h3", "h4", "h5", "h6", "cite", "code", "em", "s",
1556 "strike", "strong", "tt", "var", "div", "center",
1557 "blockquote", "ol", "ul", "dl", "table", "caption", "pre",
1558 "ruby", "rt" , "rb" , "rp", "p"
1560 $htmlsingle = array(
1561 "br", "hr", "li", "dt", "dd"
1563 $htmlnest = array( # Tags that can be nested--??
1564 "table", "tr", "td", "th", "div", "blockquote", "ol", "ul",
1565 "dl", "font", "big", "small", "sub", "sup"
1567 $tabletags = array( # Can only appear inside table
1571 $htmlpairs = array();
1572 $htmlsingle = array();
1573 $htmlnest = array();
1574 $tabletags = array();
1577 $htmlsingle = array_merge( $tabletags, $htmlsingle );
1578 $htmlelements = array_merge( $htmlsingle, $htmlpairs );
1580 $htmlattrs = $this->getHTMLattrs () ;
1582 # Remove HTML comments
1583 $text = preg_replace( "/(\\n *<!--.*--> *(?=\\n)|<!--.*-->)/sU", "$2", $text );
1585 $bits = explode( "<", $text );
1586 $text = array_shift( $bits );
1588 $tagstack = array(); $tablestack = array();
1589 foreach ( $bits as $x ) {
1590 $prev = error_reporting( E_ALL
& ~
( E_NOTICE | E_WARNING
) );
1591 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1593 list( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1594 error_reporting( $prev );
1597 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1601 if ( ! in_array( $t, $htmlsingle ) &&
1602 ( count($tagstack) && $ot = array_pop( $tagstack ) ) != $t ) {
1603 if(!empty($ot)) array_push( $tagstack, $ot );
1606 if ( $t == "table" ) {
1607 $tagstack = array_pop( $tablestack );
1612 # Keep track for later
1613 if ( in_array( $t, $tabletags ) &&
1614 ! in_array( "table", $tagstack ) ) {
1616 } else if ( in_array( $t, $tagstack ) &&
1617 ! in_array ( $t , $htmlnest ) ) {
1619 } else if ( ! in_array( $t, $htmlsingle ) ) {
1620 if ( $t == "table" ) {
1621 array_push( $tablestack, $tagstack );
1622 $tagstack = array();
1624 array_push( $tagstack, $t );
1626 # Strip non-approved attributes from the tag
1627 $newparams = $this->fixTagAttributes($params);
1631 $rest = str_replace( ">", ">", $rest );
1632 $text .= "<$slash$t $newparams$brace$rest";
1636 $text .= "<" . str_replace( ">", ">", $x);
1638 # Close off any remaining tags
1639 while ( $t = array_pop( $tagstack ) ) {
1641 if ( $t == "table" ) { $tagstack = array_pop( $tablestack ); }
1644 # this might be possible using tidy itself
1645 foreach ( $bits as $x ) {
1646 preg_match( "/^(\\/?)(\\w+)([^>]*)(\\/{0,1}>)([^<]*)$/",
1648 @list
( $qbar, $slash, $t, $params, $brace, $rest ) = $regs;
1649 if ( in_array( $t = strtolower( $t ), $htmlelements ) ) {
1650 $newparams = $this->fixTagAttributes($params);
1651 $rest = str_replace( ">", ">", $rest );
1652 $text .= "<$slash$t $newparams$brace$rest";
1654 $text .= "<" . str_replace( ">", ">", $x);
1658 wfProfileOut( $fname );
1665 * This function accomplishes several tasks:
1666 * 1) Auto-number headings if that option is enabled
1667 * 2) Add an [edit] link to sections for logged in users who have enabled the option
1668 * 3) Add a Table of contents on the top for users who have enabled the option
1669 * 4) Auto-anchor headings
1671 * It loops through all headlines, collects the necessary data, then splits up the
1672 * string and re-inserts the newly formatted headlines.
1676 /* private */ function formatHeadings( $text )
1678 global $wgInputEncoding;
1680 $doNumberHeadings = $this->mOptions
->getNumberHeadings();
1681 $doShowToc = $this->mOptions
->getShowToc();
1682 if( !$this->mTitle
->userCanEdit() ) {
1684 $rightClickHack = 0;
1686 $showEditLink = $this->mOptions
->getEditSection();
1687 $rightClickHack = $this->mOptions
->getEditSectionOnRightClick();
1690 # Inhibit editsection links if requested in the page
1691 $esw =& MagicWord
::get( MAG_NOEDITSECTION
);
1692 if( $esw->matchAndRemove( $text ) ) {
1695 # if the string __NOTOC__ (not case-sensitive) occurs in the HTML,
1697 $mw =& MagicWord
::get( MAG_NOTOC
);
1698 if( $mw->matchAndRemove( $text ) ) {
1702 # never add the TOC to the Main Page. This is an entry page that should not
1703 # be more than 1-2 screens large anyway
1704 if( $this->mTitle
->getPrefixedText() == wfMsg("mainpage") ) {
1708 # Get all headlines for numbering them and adding funky stuff like [edit]
1709 # links - this is for later, but we need the number of headlines right now
1710 $numMatches = preg_match_all( "/<H([1-6])(.*?" . ">)(.*?)<\/H[1-6]>/i", $text, $matches );
1712 # if there are fewer than 4 headlines in the article, do not show TOC
1713 if( $numMatches < 4 ) {
1717 # if the string __FORCETOC__ (not case-sensitive) occurs in the HTML,
1718 # override above conditions and always show TOC
1719 $mw =& MagicWord
::get( MAG_FORCETOC
);
1720 if ($mw->matchAndRemove( $text ) ) {
1725 # We need this to perform operations on the HTML
1726 $sk =& $this->mOptions
->getSkin();
1731 # Ugh .. the TOC should have neat indentation levels which can be
1732 # passed to the skin functions. These are determined here
1737 $sublevelCount = array();
1740 foreach( $matches[3] as $headline ) {
1743 $prevlevel = $level;
1745 $level = $matches[1][$headlineCount];
1746 if( ( $doNumberHeadings ||
$doShowToc ) && $prevlevel && $level > $prevlevel ) {
1747 # reset when we enter a new level
1748 $sublevelCount[$level] = 0;
1749 $toc .= $sk->tocIndent( $level - $prevlevel );
1750 $toclevel +
= $level - $prevlevel;
1752 if( ( $doNumberHeadings ||
$doShowToc ) && $level < $prevlevel ) {
1753 # reset when we step back a level
1754 $sublevelCount[$level+
1]=0;
1755 $toc .= $sk->tocUnindent( $prevlevel - $level );
1756 $toclevel -= $prevlevel - $level;
1758 # count number of headlines for each level
1759 @$sublevelCount[$level]++
;
1760 if( $doNumberHeadings ||
$doShowToc ) {
1762 for( $i = 1; $i <= $level; $i++
) {
1763 if( !empty( $sublevelCount[$i] ) ) {
1767 $numbering .= $sublevelCount[$i];
1773 # The canonized header is a version of the header text safe to use for links
1774 # Avoid insertion of weird stuff like <math> by expanding the relevant sections
1775 $canonized_headline = $this->unstrip( $headline, $this->mStripState
);
1778 $canonized_headline = preg_replace( "/<.*?" . ">/","",$canonized_headline );
1779 $tocline = trim( $canonized_headline );
1780 $canonized_headline = preg_replace("/[ \\?&\\/<>\\(\\)\\[\\]=,+']+/", '_', urlencode( do_html_entity_decode( $tocline, ENT_COMPAT
, $wgInputEncoding ) ) );
1781 $refer[$headlineCount] = $canonized_headline;
1783 # count how many in assoc. array so we can track dupes in anchors
1784 @$refers[$canonized_headline]++
;
1785 $refcount[$headlineCount]=$refers[$canonized_headline];
1787 # Prepend the number to the heading text
1789 if( $doNumberHeadings ||
$doShowToc ) {
1790 $tocline = $numbering . " " . $tocline;
1792 # Don't number the heading if it is the only one (looks silly)
1793 if( $doNumberHeadings && count( $matches[3] ) > 1) {
1794 # the two are different if the line contains a link
1795 $headline=$numbering . " " . $headline;
1799 # Create the anchor for linking from the TOC to the section
1800 $anchor = $canonized_headline;
1801 if($refcount[$headlineCount] > 1 ) {
1802 $anchor .= "_" . $refcount[$headlineCount];
1805 $toc .= $sk->tocLine($anchor,$tocline,$toclevel);
1807 if( $showEditLink ) {
1808 if ( empty( $head[$headlineCount] ) ) {
1809 $head[$headlineCount] = "";
1811 $head[$headlineCount] .= $sk->editSectionLink($headlineCount+
1);
1814 # Add the edit section span
1815 if( $rightClickHack ) {
1816 $headline = $sk->editSectionScript($headlineCount+
1,$headline);
1819 # give headline the correct <h#> tag
1820 @$head[$headlineCount] .= "<a name=\"$anchor\"></a><h".$level.$matches[2][$headlineCount] .$headline."</h".$level.">";
1826 $toclines = $headlineCount;
1827 $toc .= $sk->tocUnindent( $toclevel );
1828 $toc = $sk->tocTable( $toc );
1831 # split up and insert constructed headlines
1833 $blocks = preg_split( "/<H[1-6].*?" . ">.*?<\/H[1-6]>/i", $text );
1836 foreach( $blocks as $block ) {
1837 if( $showEditLink && $headlineCount > 0 && $i == 0 && $block != "\n" ) {
1838 # This is the [edit] link that appears for the top block of text when
1839 # section editing is enabled
1841 # Disabled because it broke block formatting
1842 # For example, a bullet point in the top line
1843 # $full .= $sk->editSectionLink(0);
1846 if( $doShowToc && !$i) {
1847 # Top anchor now in skin
1851 if( !empty( $head[$i] ) ) {
1860 /* private */ function doMagicISBN( &$tokenizer )
1864 # Check whether next token is a text token
1865 # If yes, fetch it and convert the text into a
1866 # Special::BookSources link
1867 $token = $tokenizer->previewToken();
1868 while ( $token["type"] == "" )
1870 $tokenizer->nextToken();
1871 $token = $tokenizer->previewToken();
1873 if ( $token["type"] == "text" )
1875 $token = $tokenizer->nextToken();
1876 $x = $token["text"];
1877 $valid = "0123456789-ABCDEFGHIJKLMNOPQRSTUVWXYZ";
1879 $isbn = $blank = "" ;
1880 while ( " " == $x{0} ) {
1882 $x = substr( $x, 1 );
1884 while ( strstr( $valid, $x{0} ) != false ) {
1886 $x = substr( $x, 1 );
1888 $num = str_replace( "-", "", $isbn );
1889 $num = str_replace( " ", "", $num );
1892 $text = "ISBN $blank$x";
1894 $titleObj = Title
::makeTitle( NS_SPECIAL
, "Booksources" );
1895 $text = "<a href=\"" .
1896 $titleObj->escapeLocalUrl( "isbn={$num}" ) .
1897 "\" class=\"internal\">ISBN $isbn</a>";
1905 /* private */ function doMagicRFC( &$tokenizer )
1909 # Check whether next token is a text token
1910 # If yes, fetch it and convert the text into a
1911 # link to an RFC source
1912 $token = $tokenizer->previewToken();
1913 while ( $token["type"] == "" )
1915 $tokenizer->nextToken();
1916 $token = $tokenizer->previewToken();
1918 if ( $token["type"] == "text" )
1920 $token = $tokenizer->nextToken();
1921 $x = $token["text"];
1922 $valid = "0123456789";
1924 $rfc = $blank = "" ;
1925 while ( " " == $x{0} ) {
1927 $x = substr( $x, 1 );
1929 while ( strstr( $valid, $x{0} ) != false ) {
1931 $x = substr( $x, 1 );
1935 $text .= "RFC $blank$x";
1937 $url = wfmsg( "rfcurl" );
1938 $url = str_replace( "$1", $rfc, $url);
1939 $sk =& $this->mOptions
->getSkin();
1940 $la = $sk->getExternalLinkAttributes( $url, "RFC {$rfc}" );
1941 $text = "<a href='{$url}'{$la}>RFC {$rfc}</a>{$x}";
1949 function preSaveTransform( $text, &$title, &$user, $options, $clearState = true )
1951 $this->mOptions
= $options;
1952 $this->mTitle
=& $title;
1953 $this->mOutputType
= OT_WIKI
;
1955 if ( $clearState ) {
1956 $this->clearState();
1959 $stripState = false;
1963 $text = str_replace(array_keys($pairs), array_values($pairs), $text);
1966 "/<br.+(clear|break)=[\"']?(all|both)[\"']?\\/?>/i" => '<br style="clear:both;"/>',
1967 "/<br *?>/i" => "<br/>",
1969 $text = preg_replace(array_keys($pairs), array_values($pairs), $text);
1970 $text = $this->strip( $text, $stripState, false );
1971 $text = $this->pstPass2( $text, $user );
1972 $text = $this->unstrip( $text, $stripState );
1976 /* private */ function pstPass2( $text, &$user )
1978 global $wgLang, $wgLocaltimezone, $wgCurParser;
1980 # Variable replacement
1981 # Because mOutputType is OT_WIKI, this will only process {{subst:xxx}} type tags
1982 $text = $this->replaceVariables( $text );
1986 $n = $user->getName();
1987 $k = $user->getOption( "nickname" );
1988 if ( "" == $k ) { $k = $n; }
1989 if(isset($wgLocaltimezone)) {
1990 $oldtz = getenv("TZ"); putenv("TZ=$wgLocaltimezone");
1992 /* Note: this is an ugly timezone hack for the European wikis */
1993 $d = $wgLang->timeanddate( date( "YmdHis" ), false ) .
1994 " (" . date( "T" ) . ")";
1995 if(isset($wgLocaltimezone)) putenv("TZ=$oldtz");
1997 $text = preg_replace( "/~~~~~/", $d, $text );
1998 $text = preg_replace( "/~~~~/", "[[" . $wgLang->getNsText(
1999 Namespace::getUser() ) . ":$n|$k]] $d", $text );
2000 $text = preg_replace( "/~~~/", "[[" . $wgLang->getNsText(
2001 Namespace::getUser() ) . ":$n|$k]]", $text );
2003 # Context links: [[|name]] and [[name (context)|]]
2005 $tc = "[&;%\\-,.\\(\\)' _0-9A-Za-z\\/:\\x80-\\xff]";
2006 $np = "[&;%\\-,.' _0-9A-Za-z\\/:\\x80-\\xff]"; # No parens
2007 $namespacechar = '[ _0-9A-Za-z\x80-\xff]'; # Namespaces can use non-ascii!
2008 $conpat = "/^({$np}+) \\(({$tc}+)\\)$/";
2010 $p1 = "/\[\[({$np}+) \\(({$np}+)\\)\\|]]/"; # [[page (context)|]]
2011 $p2 = "/\[\[\\|({$tc}+)]]/"; # [[|page]]
2012 $p3 = "/\[\[($namespacechar+):({$np}+)\\|]]/"; # [[namespace:page|]]
2013 $p4 = "/\[\[($namespacechar+):({$np}+) \\(({$np}+)\\)\\|]]/";
2014 # [[ns:page (cont)|]]
2016 $t = $this->mTitle
->getText();
2017 if ( preg_match( $conpat, $t, $m ) ) {
2020 $text = preg_replace( $p4, "[[\\1:\\2 (\\3)|\\2]]", $text );
2021 $text = preg_replace( $p1, "[[\\1 (\\2)|\\1]]", $text );
2022 $text = preg_replace( $p3, "[[\\1:\\2|\\2]]", $text );
2024 if ( "" == $context ) {
2025 $text = preg_replace( $p2, "[[\\1]]", $text );
2027 $text = preg_replace( $p2, "[[\\1 ({$context})|\\1]]", $text );
2031 $mw =& MagicWord::get( MAG_SUBST );
2032 $wgCurParser = $this->fork();
2033 $text = $mw->substituteCallback( $text, "wfBraceSubstitution" );
2034 $this->merge( $wgCurParser );
2037 # Trim trailing whitespace
2038 # MAG_END (__END__) tag allows for trailing
2039 # whitespace to be deliberately included
2040 $text = rtrim( $text );
2041 $mw =& MagicWord
::get( MAG_END
);
2042 $mw->matchAndRemove( $text );
2047 # Set up some variables which are usually set up in parse()
2048 # so that an external function can call some class members with confidence
2049 function startExternalParse( &$title, $options, $outputType, $clearState = true )
2051 $this->mTitle
=& $title;
2052 $this->mOptions
= $options;
2053 $this->mOutputType
= $outputType;
2054 if ( $clearState ) {
2055 $this->clearState();
2059 function transformMsg( $text, $options ) {
2061 static $executing = false;
2063 # Guard against infinite recursion
2069 $this->mTitle
= $wgTitle;
2070 $this->mOptions
= $options;
2071 $this->mOutputType
= OT_MSG
;
2072 $this->clearState();
2073 $text = $this->replaceVariables( $text );
2082 var $mText, $mLanguageLinks, $mCategoryLinks, $mContainsOldMagic;
2084 function ParserOutput( $text = "", $languageLinks = array(), $categoryLinks = array(),
2085 $containsOldMagic = false )
2087 $this->mText
= $text;
2088 $this->mLanguageLinks
= $languageLinks;
2089 $this->mCategoryLinks
= $categoryLinks;
2090 $this->mContainsOldMagic
= $containsOldMagic;
2093 function getText() { return $this->mText
; }
2094 function getLanguageLinks() { return $this->mLanguageLinks
; }
2095 function getCategoryLinks() { return $this->mCategoryLinks
; }
2096 function containsOldMagic() { return $this->mContainsOldMagic
; }
2097 function setText( $text ) { return wfSetVar( $this->mText
, $text ); }
2098 function setLanguageLinks( $ll ) { return wfSetVar( $this->mLanguageLinks
, $ll ); }
2099 function setCategoryLinks( $cl ) { return wfSetVar( $this->mCategoryLinks
, $cl ); }
2100 function setContainsOldMagic( $com ) { return wfSetVar( $this->mContainsOldMagic
, $com ); }
2102 function merge( $other ) {
2103 $this->mLanguageLinks
= array_merge( $this->mLanguageLinks
, $other->mLanguageLinks
);
2104 $this->mCategoryLinks
= array_merge( $this->mCategoryLinks
, $this->mLanguageLinks
);
2105 $this->mContainsOldMagic
= $this->mContainsOldMagic ||
$other->mContainsOldMagic
;
2112 # All variables are private
2113 var $mUseTeX; # Use texvc to expand <math> tags
2114 var $mUseCategoryMagic; # Treat [[Category:xxxx]] tags specially
2115 var $mUseDynamicDates; # Use $wgDateFormatter to format dates
2116 var $mInterwikiMagic; # Interlanguage links are removed and returned in an array
2117 var $mAllowExternalImages; # Allow external images inline
2118 var $mSkin; # Reference to the preferred skin
2119 var $mDateFormat; # Date format index
2120 var $mEditSection; # Create "edit section" links
2121 var $mEditSectionOnRightClick; # Generate JavaScript to edit section on right click
2122 var $mNumberHeadings; # Automatically number headings
2123 var $mShowToc; # Show table of contents
2125 function getUseTeX() { return $this->mUseTeX
; }
2126 function getUseCategoryMagic() { return $this->mUseCategoryMagic
; }
2127 function getUseDynamicDates() { return $this->mUseDynamicDates
; }
2128 function getInterwikiMagic() { return $this->mInterwikiMagic
; }
2129 function getAllowExternalImages() { return $this->mAllowExternalImages
; }
2130 function getSkin() { return $this->mSkin
; }
2131 function getDateFormat() { return $this->mDateFormat
; }
2132 function getEditSection() { return $this->mEditSection
; }
2133 function getEditSectionOnRightClick() { return $this->mEditSectionOnRightClick
; }
2134 function getNumberHeadings() { return $this->mNumberHeadings
; }
2135 function getShowToc() { return $this->mShowToc
; }
2137 function setUseTeX( $x ) { return wfSetVar( $this->mUseTeX
, $x ); }
2138 function setUseCategoryMagic( $x ) { return wfSetVar( $this->mUseCategoryMagic
, $x ); }
2139 function setUseDynamicDates( $x ) { return wfSetVar( $this->mUseDynamicDates
, $x ); }
2140 function setInterwikiMagic( $x ) { return wfSetVar( $this->mInterwikiMagic
, $x ); }
2141 function setAllowExternalImages( $x ) { return wfSetVar( $this->mAllowExternalImages
, $x ); }
2142 function setSkin( $x ) { return wfSetRef( $this->mSkin
, $x ); }
2143 function setDateFormat( $x ) { return wfSetVar( $this->mDateFormat
, $x ); }
2144 function setEditSection( $x ) { return wfSetVar( $this->mEditSection
, $x ); }
2145 function setEditSectionOnRightClick( $x ) { return wfSetVar( $this->mEditSectionOnRightClick
, $x ); }
2146 function setNumberHeadings( $x ) { return wfSetVar( $this->mNumberHeadings
, $x ); }
2147 function setShowToc( $x ) { return wfSetVar( $this->mShowToc
, $x ); }
2149 /* static */ function newFromUser( &$user )
2151 $popts = new ParserOptions
;
2152 $popts->initialiseFromUser( $user );
2156 function initialiseFromUser( &$userInput )
2158 global $wgUseTeX, $wgUseCategoryMagic, $wgUseDynamicDates, $wgInterwikiMagic, $wgAllowExternalImages;
2160 if ( !$userInput ) {
2162 $user->setLoaded( true );
2164 $user =& $userInput;
2167 $this->mUseTeX
= $wgUseTeX;
2168 $this->mUseCategoryMagic
= $wgUseCategoryMagic;
2169 $this->mUseDynamicDates
= $wgUseDynamicDates;
2170 $this->mInterwikiMagic
= $wgInterwikiMagic;
2171 $this->mAllowExternalImages
= $wgAllowExternalImages;
2172 $this->mSkin
=& $user->getSkin();
2173 $this->mDateFormat
= $user->getOption( "date" );
2174 $this->mEditSection
= $user->getOption( "editsection" );
2175 $this->mEditSectionOnRightClick
= $user->getOption( "editsectiononrightclick" );
2176 $this->mNumberHeadings
= $user->getOption( "numberheadings" );
2177 $this->mShowToc
= $user->getOption( "showtoc" );
2183 # Regex callbacks, used in Parser::replaceVariables
2184 function wfBraceSubstitution( $matches )
2186 global $wgCurParser;
2187 return $wgCurParser->braceSubstitution( $matches );